2024 · Case study

TrinetraAI

Hands-free desktop control for users who can't rely on keyboard and mouse.

Built a real-time gesture pipeline mapping webcam input to OS-level actions.

The problem

Conventional desktop interfaces assume full keyboard and mouse access — leaving users with limited mobility dependent on expensive assistive hardware.

Why it's interesting

Turns a commodity webcam into an accessibility interface using real-time computer vision, not specialized hardware.

What I built

Webcam-based gesture recognition pipeline
Gesture-to-action mapping for desktop control
Real-time inference loop with OpenCV + MediaPipe

Technical components

PythonOpenCVMediaPipe hand trackingReal-time video processingDesktop automation hooks

What made it non-trivial

Low-latency gesture detection that remains reliable across varying lighting and camera angles.

My role

Solo engineer — computer vision, gesture mapping, desktop integration.