Inspiration
Hand tremors affect millions of people with conditions like Parkinson’s disease and essential tremor, making simple daily tasks difficult. Most existing solutions are either expensive, bulky, or inaccessible. We wanted to explore whether modern computer vision and web technologies could be used to visualize and digitally reduce tremors in real time, without requiring specialized hardware.
What it does
Stabil-Arm is a real-time hand-tracking and tremor visualization system. It captures live hand movement through a standard camera, detects tremors, and applies a stabilization algorithm to show a smoother, controlled virtual arm movement. The system helps demonstrate how assistive stabilization can improve precision and control.
How we built it
We used MediaPipe Hands for real-time hand landmark detection and built the frontend using React.js. The raw hand movement data is processed using custom smoothing and filtering logic to reduce tremor intensity. The stabilized motion is then visualized through an interactive web interface using Canvas and 3D rendering techniques. Everything runs directly in the browser, requiring no external hardware.
Challenges we ran into
The biggest challenge was balancing responsiveness and stability. Over-smoothing caused noticeable lag, while under-smoothing failed to reduce tremors effectively. Achieving real-time performance in the browser while keeping the visualization accurate required multiple iterations and tuning of the stabilization logic.
What we learned
We learned how powerful browser-based computer vision has become and how assistive technology can be prototyped quickly using web tools. This project deepened our understanding of real-time signal smoothing, human–computer interaction, and performance optimization in frontend systems.
What’s next
Future versions could integrate hardware actuators, adaptive AI-based tremor prediction, and clinical feedback loops. The long-term goal is to move from visualization to a fully functional assistive device.
Log in or sign up for Devpost to join the conversation.