👀 Eye Disease Detection - Deep Learning Flask App

This is an AI-powered web application for detecting eye diseases from retinal images. It uses a custom-trained CNN based on the VGG16 architecture.

---

🔗 Live Demo

👉 Click here to try the app
Hosted on Render — no installation needed.

---

🔍 Supported Eye Conditions

• 👁 Cataract
• 🩸 Diabetic Retinopathy
• 🌫 Glaucoma
• ✅ Normal
• ❓ Unknown (for cases the model can't classify confidently)

---

🚀 How It Works

1. Upload a retina image from your local system.
2. The image is preprocessed and fed into a CNN.
3. The model outputs the predicted class and confidence score.
4. The result is shown visually in the browser.

---

🧠 Model Info

• Architecture: Transfer learning using VGG16
• Input Size: 224 x 224
• Framework: TensorFlow + Keras
• Classes: ["Unknown", "Cataract", "Diabetic Retinopathy", "Glaucoma", "Normal"]

---

📁 File Structure

.
├── app.py                             # Flask web app
├── templates/
│   ├── index.html                     # Upload page
│   ├── result.html                    # Result page
│   └── partials/
│       ├── _disease_info.html         # Info about eye diseases
│       └── _how_it_works.html         # Model explanation section
├── static/
│   ├── styles.css                     # Custom styles
│   └── js/
│       └── main.js                    # JavaScript for interactivity
├── models/
│   ├── eye_disease_model.h5          # Trained model (downloaded automatically)
│   └── eye_diseases_detection_vgg16.ipynb # Model training notebook
└── requirements.txt                  # Required Python packages

---

📦 Requirements

• Python 3.8+
• TensorFlow
• Flask
• Pillow
• NumPy
• gdown (for Google Drive file download)

Install all dependencies:

pip install -r requirements.txt

---

📥 Model Auto-Download

If models/eye_disease_model.h5 is not found, the app automatically downloads it from:

👉 https://drive.google.com/file/d/11nt9ilopu9RmYIjCeAhOjaQbaREG3PG-/view

---

🖥 Run the App

python app.py

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

---

🤝 Contribution & Contact

This project is open for contributions! If you have suggestions, fixes, or ideas for improvement, feel free to open an issue or PR.

• 📧 Email: mdiibrahim549@gmail.com
• 💼 LinkedIn: Mohammad Ibrahim

---

⚖️ License

This project is licensed under the MIT License.