🔬 Diabetic Retinopathy Detection System

An advanced AI-powered web application for detecting and classifying diabetic retinopathy from retinal fundus images using deep learning.

🌟 Features

• AI-Powered Detection: Uses ResNet-152 deep learning architecture for accurate classification
• 5 Severity Levels: Classifies diabetic retinopathy into 5 stages (No DR, Mild, Moderate, Severe, Proliferative DR)
• Modern Web Interface: Beautiful, responsive UI with real-time analysis
• High Accuracy: Trained model achieves medical-grade accuracy
• Detailed Results: Provides confidence scores, probability distributions, and medical recommendations
• Easy to Use: Simple drag-and-drop interface for image upload

📋 Table of Contents

• Features
• Installation
• Model Setup
• Usage
• API Documentation
• Project Structure
• Technology Stack
• Research Paper
• License

🚀 Installation

Prerequisites

• Python 3.8 or higher
• pip package manager
• 4GB+ RAM recommended
• CUDA-capable GPU (optional, for faster inference)

Step 1: Clone the Repository

git clone <repository-url>
cd Diabetic-Retinopathy-Detection-main

Step 2: Create Virtual Environment (Recommended)

# Windows
python -m venv venv
venv\Scripts\activate

# Linux/Mac
python3 -m venv venv
source venv/bin/activate

Step 3: Install Dependencies

pip install -r requirements.txt

🧠 Model Setup

The trained model file (classifier.pt) needs to be downloaded separately due to its large size.

Download the Model

1. Download the pre-trained ResNet-152 model from:

   • Kaggle: Blindness Detection Pretrained Weights

2. Place the classifier.pt file in the following location:

   Diabetic-Retinopathy-Detection-main/
   └── Retinal_blindness_detection_Pytorch-master/
       └── classifier.pt
   

Verify Model Path

Ensure the model path in app.py matches your setup:

MODEL_PATH = 'Retinal_blindness_detection_Pytorch-master/classifier.pt'

💻 Usage

Starting the Application

1. Start the Backend Server:

   python app.py

   The server will start on http://localhost:5000

2. Access the Web Interface:

   • Open your browser and navigate to: http://localhost:5000
   • Or open frontend/index.html directly in your browser

Using the Application

1. Upload Image:

   • Click on the upload area or drag and drop a retinal fundus image
   • Supported formats: PNG, JPG, JPEG (Max 16MB)

2. Analyze:

   • Click the "Analyze Image" button
   • Wait for the AI model to process the image

3. View Results:

   • See the severity classification
   • Review confidence scores and probability distribution
   • Read medical recommendations

Sample Images

Test images are available in:

Retinal_blindness_detection_Pytorch-master/sampleimages/

📡 API Documentation

Base URL

http://localhost:5000/api

Endpoints

1. Health Check

GET /api/health

Response:

{
  "status": "healthy",
  "model_loaded": true,
  "device": "cuda" | "cpu"
}

2. Predict

POST /api/predict

Request:

• Content-Type: multipart/form-data
• Body: file (image file)

Response:

{
  "severity_value": 0,
  "severity_class": "No DR",
  "confidence": 98.5,
  "probabilities": {
    "No DR": 98.5,
    "Mild": 1.2,
    "Moderate": 0.2,
    "Severe": 0.05,
    "Proliferative DR": 0.05
  },
  "info": {
    "level": "No DR",
    "description": "No signs of diabetic retinopathy detected.",
    "recommendation": "Continue regular eye examinations...",
    "color": "#10b981",
    "risk": "Low"
  },
  "image_data": "data:image/jpeg;base64,..."
}

3. Get Classes

GET /api/classes

Response:

{
  "classes": ["No DR", "Mild", "Moderate", "Severe", "Proliferative DR"],
  "info": [...]
}

📁 Project Structure

Diabetic-Retinopathy-Detection-main/
├── app.py                          # Flask backend server
├── requirements.txt                # Python dependencies
├── README.md                       # This file
├── LICENSE                         # License file
├── Research Paper.pdf              # Published research paper
│
├── frontend/                       # Web frontend
│   ├── index.html                  # Main HTML file
│   ├── styles.css                  # Styling
│   └── script.js                   # JavaScript logic
│
├── Retinal_blindness_detection_Pytorch-master/
│   ├── classifier.pt               # Trained model (download separately)
│   ├── model.py                    # Model architecture
│   ├── blindness.py                # Original GUI application
│   ├── sampleimages/               # Test images
│   ├── training.ipynb              # Training notebook
│   ├── inference.ipynb             # Inference notebook
│   └── requirements.txt            # Original requirements
│
└── uploads/                        # Temporary upload folder (auto-created)

🛠️ Technology Stack

Backend

• Framework: Flask 2.3.3
• Deep Learning: PyTorch 1.13.1
• Model: ResNet-152 (pretrained and fine-tuned)
• Image Processing: Pillow, torchvision
• API: RESTful API with CORS support

Frontend

• HTML5: Semantic markup
• CSS3: Modern styling with gradients, animations, glassmorphism
• JavaScript: Vanilla JS (no frameworks)
• Fonts: Inter, Space Grotesk (Google Fonts)

Model Architecture

• Base Model: ResNet-152
• Custom Classifier:
  • Linear(2048 → 512) + ReLU
  • Linear(512 → 5) + LogSoftmax
• Loss Function: Negative Log Likelihood Loss
• Optimizer: Adam
• Input Size: 224×224 RGB images

📊 Severity Levels

Level	Classification	Description	Risk
0	No DR	No signs of diabetic retinopathy	Low
1	Mild	Mild non-proliferative diabetic retinopathy	Low-Medium
2	Moderate	Moderate non-proliferative diabetic retinopathy	Medium
3	Severe	Severe non-proliferative diabetic retinopathy	High
4	Proliferative DR	Proliferative diabetic retinopathy	Critical

📄 Research Paper

This project is based on published research on diabetic retinopathy detection using deep learning. The research paper is included in the repository as Research Paper.pdf.

🔧 Troubleshooting

Model Not Loading

• Ensure classifier.pt is in the correct location
• Check the MODEL_PATH variable in app.py
• Verify the model file is not corrupted

CORS Errors

• Ensure Flask-CORS is installed: pip install flask-cors
• Check that the backend server is running

Out of Memory

• Use CPU instead of GPU for inference
• Reduce batch size (already set to 1 for single image inference)
• Close other applications

Port Already in Use

• Change the port in app.py:

  app.run(debug=True, host='0.0.0.0', port=5001)

🤝 Contributing

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

⚠️ Disclaimer

This tool is for educational and screening purposes only. It is not a substitute for professional medical diagnosis. Always consult with a qualified healthcare professional for medical advice and treatment.

📝 License

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

👥 Authors

• Original Model & Research: [Research Team]
• Web Application: [Your Name]

🙏 Acknowledgments

• ResNet architecture by Microsoft Research
• PyTorch framework
• Kaggle for hosting the dataset and model weights
• Medical professionals for domain expertise

📧 Contact

For questions or support, please open an issue in the repository.

---

Made with ❤️ for early detection and prevention of diabetic retinopathy