Artificial Intelligence

Real-Time Arabic Sign Language Translation

An intelligent system recognizing Arabic hand signs in real time and presenting text translation.

Technologies used
  • Python
  • Machine Learning
  • Data Analysis
Real-Time Arabic Sign Language Translation

Project overview

The system was built by collecting and organizing images of hand gestures, extracting 63 features from 21 key hand points using MediaPipe, and then training a neural model to classify the gestures. The system supports dozens of categories and uses a real-time camera to display the predicted gesture and Arabic text in a way that facilitates communication, education, and scientific research.

Key features

  • Over 15,000 original images.
  • A processing dataset containing over 19,000 records.
  • Support for 69 signal categories.
  • Hand feature extraction from key points.
  • Approximately 94.59% evaluation accuracy.
All projects

Discuss your project with us

Send us your idea or requirements so we can review the scope and agree on suitable next steps.

Contact