🏥 Diabetic Retinopathy Detection using Ensemble AI Approach

📌 Overview

This project leverages an Ensemble Approach for detecting Diabetic Retinopathy (DR) from retinal images. By combining multiple pretrained CNN models, the system enhances diagnostic accuracy and robustness for detection of DR.

🚀 Features:

✅ Ensemble Model – Utilizes multiple CNN architectures for superior performance.
✅ High Accuracy – Optimized for real-world clinical settings.
✅ Flask Web App – User-friendly interface for easy image uploads and predictions.
✅ Scalable Deployment – Can be hosted locally or on cloud platforms.

🖼️ Model Performance

Above: An example of the ensemble model's performance on test data.

⚙️ Technologies Used

Technology	Logo
Python
PyTorch
TensorFlow
Flask
OpenCV
Matplotlib

📂 Project Structure

📁 diabetic-retinopathy-detection
│-- app.py                 # Flask application
│-- model.py               # Ensemble AI model (CNN-based)
│-- static/
│   ├── uploads/           # Uploaded images
│   ├── results/           # Processed images
│-- templates/
│   ├── index.html         # Main upload page
│   ├── result.html        # Prediction page
│-- requirements.txt       # Dependencies
│-- README.md              # Project documentation

🎯 How to Run

1️⃣ Clone the Repository

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

2️⃣ Install Dependencies

pip install -r requirements.txt

3️⃣ Run the Flask Server

python app.py

4️⃣ Open in Browser

Visit http://127.0.0.1:5000 in your web browser.

📌 Example Usage

1️⃣ Upload an eye fundus image.
2️⃣ The ensemble model predicts the Diabetic Retinopathy stage.
3️⃣ Results are displayed with confidence scores.

🎨 UI Preview

📊 Results

Retinopathy Stage	Model Prediction	Confidence
Mild DR	✅ Correct	94.7%
Proliferative DR	✅ Correct	96.3%

📖 Future Improvements

• Improve dataset augmentation for better generalization.
• Optimize inference speed for real-time applications.
• Deploy on cloud platforms like AWS/Google Cloud.

🤝 Acknowledgments

Special thanks to the medical and AI communities for their valuable datasets and research.
Inspirational guidance from Dr. Victor Ikechukwu. Explore their work: Dr. Victor Ikechukwu.

📜 License

This project is licensed under the MIT License.

---

🔥 If you like this project, don't forget to ⭐ it on GitHub!