nnfl-project

Stars
6
Language
Jupyter Notebook
License
MIT
Updated
View on GitHub

Real-Time Sign Language Recognition Using Neural Networks

Author: Manav Arya and Kiritika Prakash

This project is solely owned and maintained by Manav Arya. For more projects, visit my GitHub: nnfl-project

Overview

This project provides a real-time sign language detection system using an LSTM (Long Short-Term Memory) neural network. It enables users to recognize sign language gestures from video input, making communication more accessible for the hearing and speech impaired.

Features

  • Real-time detection of sign language gestures
  • Deep learning model (LSTM) for sequence prediction
  • Easy-to-use interface
  • Pre-trained model weights included

Project Structure

├── model_weights.h5         # Pre-trained model weights
├── model.h5                 # Model architecture
├── RealTimeSignLanguageDetection.ipynb  # Demo notebook
├── Train.ipynb              # Model training notebook
├── v2/
│   ├── main.py              # Main script for detection
│   ├── requirements.txt     # Python dependencies
│   └── README.md            # v2 module documentation

Setup Instructions

  1. Clone the repository: 
    git clone https://github.com/Manavarya09/Realtime-Sign-Language-Detection-Using-LSTM-Model.git
cd Realtime-Sign-Language-Detection-Using-LSTM-Model-main
  2. Create and activate a virtual environment: 
    python3 -m venv venv
source venv/bin/activate
  3. Install dependencies: 
    pip install -r v2/requirements.txt
pip install notebook

Usage

  • Run the main detection script: 
    python v2/main.py
  • Open and run the demo notebook: 
    jupyter notebook RealTimeSignLanguageDetection.ipynb

Model Training

To retrain the model, use the Train.ipynb notebook. Make sure your dataset is properly formatted and update the paths as needed.

License

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

Contact

For questions or contributions, please contact:

This project is solely owned and maintained by Manav Arya. Feel free to open issues or submit pull requests to improve this project!