DIALEX

Dialex: Connecting Hands and Voices

What Inspired Me

In my neighborhood, I often see a group of deaf and dumb individuals walking together, using their hands to communicate fluently. I didn’t understand what they were saying, but I was always intrigued by how expressive their gestures were.

Having worked on several hand tracking projects in VR and AR, I started wondering: “What if I could build something that lets people like me understand them — and lets them understand us too?”

That was the moment Dialex was born — a project to bridge the communication gap between signers and non-signers.

How I Built Dialex

Dialex is a bidirectional sign language translator: • It converts spoken words to signs, using speech-to-text and animated sign models. • And it also converts signs to text, using hand tracking and gesture recognition.

I used MediaPipe for hand detection, trained a lightweight gesture model, and built the system in Unity, combining what I knew from VR development with accessibility goals.

What I Learned • Sign language is rich — more than just waving hands. • Real-time systems need to be smooth and simple. • Accessibility pushes you to design with empathy first.

Challenges I Faced • Getting smooth sign animations for each word • Training the model to recognize similar signs • Optimizing the app for mobile devices

End Goal

Dialex isn’t just a tech demo — it’s about breaking the silence, making sure everyone can be heard and understood, no matter how they communicate.

Built With

• amazon-web-services
• api
• c#
• groq
• lite
• mediapipe
• python
• tasks
• tensorflow
• unity
• urp)