The Diabetic Retinopathy Detection project is an AI-powered solution aimed at facilitating early diagnosis of diabetic retinopathy through image-based analysis. Leveraging the power of Convolutional Neural Networks (CNN) and Deep Learning frameworks, the system analyzes retinal images to detect various stages of diabetic retinopathy, thus aiding in timely intervention and treatment.
This project uses a comprehensive dataset of retinal fundus images and implements a multi-layered CNN architecture to achieve high accuracy in identifying symptoms such as microaneurysms, hemorrhages, and exudates. The model is hosted on a web interface, enabling healthcare professionals to upload retinal images and receive instant diagnostic results.
-
🔹 Automated Diagnosis
Automatically classifies retinal images into different stages of diabetic retinopathy, minimizing the need for manual examination. -
🔹 CNN-Based Model
A well-trained Convolutional Neural Network with multiple layers for feature extraction and classification. -
🔹 Web-Based Interface
A user-friendly web interface where users can upload images, view results, and understand the diagnostic confidence levels. -
🔹 Interactive Visualizations
Generates heatmaps and overlay visualizations to highlight areas of concern in the retinal images. -
🔹 Cloud Hosting
Hosted on a cloud platform for remote access, with support for real-time data processing.
-
Programming Language:
Used for implementing the model and backend functionalities. -
Frameworks:
The primary library for building and training the deep learning model.
Used for constructing the neural network layers and model management. -
Frontend & Hosting:
Streamlit is used for building the web interface, and the application is hosted on Heroku for easy access and scalability. -
Libraries Used:
NumPyPandasOpenCVMatplotlib
Before you begin, ensure you have the following installed on your local machine:
- Python 3.8 or above
- TensorFlow and Keras libraries
- Streamlit for the web interface
Follow the steps below to set up the project on your local environment:
- Navigate to the Project Directory:
cd diabetic-retinopathy-detection - Install Dependencies: Install the necessary dependencies using the requirements.txt file
pip install -r requirements.txt
- Run the Application: Launch the Streamlit app locally.
streamlit run app.py
- Access the Interface: Open your web browser and go to the following URL to interact with the application
http://localhost:8501
The model achieves an accuracy of 85% on the validation dataset and a precision of 88%. The results are visualized using confusion matrices and ROC curves, providing a clear understanding of the model's classification performance. The model has been fine-tuned to minimize false negatives, ensuring reliable predictions and robust detection of diabetic retinopathy stages.
- 🌟 Implement Hybrid Models: Integrate a hybrid model combining CNNs with RNNs to capture temporal dependencies for more comprehensive analysis.
- 🌟 Expand Diagnostic Scope: Extend the model's capabilities to include detection of other retinal conditions such as glaucoma and macular degeneration.
- 🌟 Enhanced Web Interface: Upgrade the web interface with more interactive diagnostic insights and patient management features, making it a one-stop solution for retinal health monitoring.
Contributions are welcome! If you have suggestions for improvements or new features, feel free to open an issue or create a pull request. Let's collaborate to make this project even better!
For any inquiries or suggestions, please feel free to reach out:
- Name: Heet Mehta
- Email: mehtaheet5@gmail.com
- GitHub: Heet852003