JointScope

Front page
Dashboard
Video Recovery Test
Recommended Exercises

Inspiration

Generic rehab programs can lead to reinjury rates as high as 25%, while personalized recovery plans reduce that risk to about 4%. This inspired JointScope to move beyond time-based recovery and use real movement data to help people progress safely.

What it does

JointScope uses video and pose estimation to measure joint range of motion in real time, giving users feedback on their recovery. Based on these measurements, it recommends appropriate physical therapy exercises and tracks progress over time to ensure safe, effective recovery.

How we built it

We used MediaPipe Pose to get the angle of joint and used JavaScript to connect to camera and pose estimator and used HTML, CSS and React to create the frontend webpage.

Challenges we ran into

Integrating the HTML and CSS frontend with the JavaScript logic and MediaPipe Pose pipeline required careful coordination between real-time camera input, pose estimation. Joint angles initially updated every video frame, causing excessive fluctuation. To improve stability and usability, we adjusted the system to compute and update angles at fixed time intervals.

What we learned

We learned a lot of Git commands and version control. We learned to use React to connect to backend. We learned how to connect camera containers, biomechanics and how ML learning models assess the body.

What's next for JointScope

JointScope will evolve to be more interactive by giving users real-time feedback to help them perform exercises correctly and safely. Recovery data can be shared with physical therapists or doctors. As users improve.