Breaking communication barriers with high-performance computer vision. Our system turns live visual gestures into reliable, actionable output in real-time.
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.
95.88% recognition rate on ArSL dataset
Privacy-first: no video data ever leaves your device
Experience the translation system in action. Grant camera access to start detecting Arabic sign language letters in real-time.
High-fidelity hand and finger tracking providing 21 3D landmarks for precise gesture analysis.
Hardware-accelerated machine learning running directly in the browser via WebGL.
Custom-trained neural network optimized for the Arabic Sign Language (ArSL) alphabet.
Zero-latency inference by processing everything locally on the user's hardware.
Facilitating communication for the deaf and hard-of-hearing community in Arabic-speaking regions.
Integrating sign language support into public service kiosks, transit hubs, and government offices.
Helping students learn Arabic Sign Language with real-time feedback and interactive sessions.
Let's work together to build the future of communication and public service.
Contact Us