In the past, motion analysis for physical therapy required markers or sensors and expensive cameras. However, new AI-based pose detection technology allows for real-time, high-accuracy motion tracking with just a built-in camera, enabling improved musculoskeletal outcomes for patients.
My team developed a markerless pose detection technology to be integrated to telerehabilitation solutions used by MSK physical therapists.
Here is a prompt video how this technology tracks an exercise performance: https://www.youtube.com/shorts/uEQdF2OqqTE
We built it keeping in mind that most of the competitors reach the needed accuracy with the help of bulky sensors and special cameras.
What we wanted to reach with our technology – is to make it device-independent, sensorless yet accurate and being able to process MSK screening in real-time.
And while we're still working towards adding more complex movements to check the motion analysis process you're all welcome to evaluate the idea and share you feedback on what we've done.
p.s. In future we're about to expand the technology for wide range of MSK exercises and:
- Add age and gender normative analysis
- Automatic day-to-day progress comparison
- ML-generated exercises’ recommendation, e.g. actionable plans to improve joints’ functional status
- Asymmetries collecting with visual alerts
- Skeleton graph and measurement
Suggestions and ideas are welcome!
Top comments (0)