AI Research & Demo

Arabic Sign Language
ML Translation System

Breaking communication barriers with high-performance computer vision. Our system turns live visual gestures into reliable, actionable output in real-time.

Empowering Accessibility Through AI

We developed an Arabic Sign Language recognition system for real-world use, achieving 95.88% accuracy using MediaPipe and TensorFlow.js. This project demonstrates how edge-based AI can provide instant, private, and accessible communication tools.

High Accuracy

95.88% recognition rate on ArSL dataset

Local Processing

Privacy-first: no video data ever leaves your device

Live Demo

Experience the translation system in action. Grant camera access to start detecting Arabic sign language letters in real-time.

The Technology Stack

MediaPipe Hands

High-fidelity hand and finger tracking providing 21 3D landmarks for precise gesture analysis.

TensorFlow.js

Hardware-accelerated machine learning running directly in the browser via WebGL.

Deep Learning Model

Custom-trained neural network optimized for the Arabic Sign Language (ArSL) alphabet.

Edge Computing

Zero-latency inference by processing everything locally on the user's hardware.

Impact & Applications

Inclusion

Facilitating communication for the deaf and hard-of-hearing community in Arabic-speaking regions.

Smart Kiosks

Integrating sign language support into public service kiosks, transit hubs, and government offices.

Educational Tools

Helping students learn Arabic Sign Language with real-time feedback and interactive sessions.

Looking for custom AI solutions?

Let's work together to build the future of communication and public service.

Contact Us
Burj Khalifa
Burj Al Arab
Building 1Building 2
Horizontal Building