GestureTalk is a real-time gesture-to-speech system that converts hand gestures into spoken phrases using computer vision and AI. It is designed to assist individuals with speech impairments by enabling seamless communication in everyday environments.
People with speech or hearing impairments often struggle with real-time communication in places like hospitals, cafes, and colleges. Existing solutions are either expensive or not intuitive. This project provides a low-cost, real-time alternative using only a camera and AI-based hand tracking.
- Python
- Flask
- OpenCV
- MediaPipe
- NumPy
- pyttsx3
- HTML, CSS, JavaScript
python -m venv venv311
venv311\Scripts\activateMediaPipe may not support the latest Python versions.
👉 Use:
- Python 3.10 or 3.11
Check:
python --versionpip install -r requirements.txtpython app.pyOpen in browser:
http://localhost:5000
app.py
camera.py
gestures.py
tts.py
config.py
templates/
static/
requirements.txt
- Real-time hand gesture recognition
- Context-based phrase mapping (cafe, hospital, etc.)
- Text-to-speech output
- Custom gesture training
- Live dashboard with status updates
Large files (like virtual environments) were accidentally pushed and later removed by:
rmdir /s /q .git
git init
git branch -M main
echo venv311/ >> .gitignore
echo __pycache__/ >> .gitignore
echo *.pyc >> .gitignore
echo *.pyd >> .gitignore
echo *.dll >> .gitignore
git add .
git commit -m "Clean project upload"
git remote add origin https://github.com/Blueundead24/Assistive-technology.git
git push -u origin main --force- Mobile application version
- Multi-language speech support
- Pre-trained gesture dataset
- Cloud deployment
- Added stable gesture capture before training
- Fixed unreliable live training issue
- Improved camera handling (webcam support)