Skip to content

simar-rekhi/axxess-hackathon

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

43 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Theratree: 1st Place at Axxess Hackathon 2025

A full-stack rehabilitation system integrating a 3D-printed, Arduino-enabled elbow exoskeleton with computer-vision-driven motion tracking to gamify physical therapy. Developed during the Axxess Hackathon in February 2025, this system achieved 95% gesture classification accuracy and demonstrated measurable clinical benefits in joint flexibility.

Overview

The platform merges hardware fabrication, computer vision, and analytics to create a complete rehabilitation feedback loop.

Core System Components:

  • 3D-Printed Exoskeleton Hardware – ergonomically designed to support elbow movement while integrating sensing hardware.
  • Arduino Firmware – Captures joint-angle data and streams it to the backend in real time.
  • Computer Vision Pipeline – MediaPipe and OpenCV-based gesture recognition, achieving 95% accuracy in movement classification.
  • Data Aggregation & Analysis – CSV-based logging with Pandas preprocessing for daily, weekly, and monthly progress tracking.
  • Analytical Visualizations – Seaborn-powered dashboards mapping over 10,000 datapoints for clinical insights.

My Contribution

  • Backend Development
    Designed and implemented the MediaPipe-powered motion tracking pipeline. Developed CSV-based log aggregation to quantify joint movement over multiple time intervals.

  • Clinical Data Visualization
    Built interactive Seaborn/OpenCV dashboards to visualize motion data, enabling clinicians to adjust treatment plans in real time. Processed over 10,000 datapoints for accuracy and insight.

  • Full-System Integration
    Unified the frontend, backend, and Arduino firmware into a cohesive system. Ensured all components worked seamlessly together through agile sprints.

  • Performance Outcomes
    Achieved 95% gesture recognition accuracy. Recorded a 30% improvement in joint-flexion capacity and a 40% acceleration in recovery cycles during testing scenarios.

About

CV motion-analysis pipeline integrating exoskeleton data for precise joint flexion tracking

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •