🏆 This project was awarded Most Technical at FullyHacks 2025 🏆

🪐 FullyHacks 2025

FullyHacks is the largest 24-hour Hackathon at CSUF hosted by ACMCSUF

👽 ASTRO ASL

No sound in space? No Problem. Real-time ASL recognition via deep learning 💫

📝 Summary

ASTRO ASL (American Sign Language) is an AI-powered sign language transcriber designed specifically for space missions. By capturing and interpreting ASL gestures in real time, ASTRO ASL bridges communication gaps in environments where traditional audio communication fails.

🎯 Mission Statement

Our mission is to empower astronauts with a reliable, silent communication tool.
We strive to:

• Enable non-verbal communication in space where sound doesn't travel
• Support inclusivity for deaf or hard-of-hearing astronauts
• Enhance safety through clear, real-time gesture recognition

By combining cutting-edge AI with intuitive design, ASTRO ASL provides a seamless way for astronauts to interact without relying on sound — ensuring mission-critical information is never lost in translation.

⚠️ The Problem

Sound doesn’t travel in the vacuum of space, making traditional spoken communication unreliable during spacewalks or in loud environments.

• Verbal communication is limited in spacewalks
• Audio equipment can malfunction or be blocked by suits
• Safety risks increase when commands aren't clearly heard

ASTRO ASL provides a hands-on solution to a soundless environment.

✨ Features

Feature	Description
🤖 AI Recognition	Real-time ASL interpretation via onboard camera systems
🧠 Onboard ML	No internet needed — edge processing for zero-latency use
🧤 Glove Support*	Compatible with bulky astronaut gloves
📊 Scribe Logs*	Automatically logs conversations for mission review
🌌 Space-Ready*	Designed for zero-gravity and suit integration

* TBD (To Be Developed...)

🌟 How Are We Unique?

• Tech Used: TensorFlow, OpenCV, Mediapipe, Scikit-learn
• Built for Space: Engineered to work in zero-gravity and vacuum conditions
• Offline AI: Works without any internet connection, runs locally
• Mission-Critical UX: Simplified UI for high-pressure scenarios

🚀 Steps to Use

1. Clone Repository:

   git clone git@github.com:isliese/astro-asl
   cd astro-asl

2. Set Up Environment:

   python3 -m venv venv
   source venv/bin/activate  # Linux/macOS
   .\venv\Scripts\activate   # Windows

3. Install Dependencies:

   # Ensure Python > 3.10 < 3.11 is installed
   pip install -r requirements.txt

4. Train Model

   # Using your own dataset, you can train your model using:
   # Only tested using dataset below. May require customization
   # for custom datasets.
   # https://www.kaggle.com/datasets/ayuraj/american-sign-language-dataset
   python training_model.py

5. Run Application:

   python app.py # for web UI and transcriber

🔍 Troubleshooting

Issue	Solution
Performance Slowdowns	Ensure efficient use of system resources
False Positives/Negatives	Strengthen model to reduce overfitting
Installation Issues	Verify Python version and dependencies

🤝 Contributing

We welcome contributions! Please follow these steps:

1. Fork the repository
2. Create a new branch for your feature or bugfix
3. Submit a pull request with a detailed description of your changes

👥 Team

Role	Member
Backend / Research	Owin
Backend / Research	Jay
Frontend / Design	Isla
Frontend / Design	Sema

🙏 Credits

This project uses the following technologies, libraries, and datasets:

Languages:

• Python
• JS

Libraries:

• Flask – A micro web framework for Python.
• TensorFlow – An open-source library for numerical computation and large-scale machine learning.
• Scikit-learn – A machine learning library, used for building the Random Forest model.
• NumPy – A library for numerical computing in Python.
• MediaPipe – A library for real-time computer vision.

Dataset:

American Sign Language Dataset – A dataset used for training the the model.

Explore More:

Astro ASL Slideshow
Devpost Project Link

📄 License

© 2025 | AstroASL Team - Fullyhacks @ CSUF