In our 2026 QMIND Healthcare project, we used an Intel RealSense D435 camera with depth-sensing technology. It combines RGB cameras with infrared stereo depth sensors to produce a hybrid image of both colour and depth. These cameras are widely used in tasks involving facial recognition.
We are using this camera in combination with MediaPipe's pretrained models to compute real-time angles of rotation in a subject's shoulders, exploring the feasibility of such a system in orthopedic ROM assessment. This is a work in progress!
Team lead: Cait Roach
Design team: Hargun Kour, Basma Azeem, Augustine Osezua, Gavin Tan, Leif Hill, Nicholas Irons, Tharunika Gnaneshan
we used a lot of the math and code from these repos/studies, so please check them out too:
- https://arxiv.org/pdf/2310.07322
- https://www.researchgate.net/publication/322815664_Sitting_posture_assessment_using_computer_vision
- https://github.com/ashita03/Bone-Fracture-Detection
- https://pitthexai.github.io/research.html
- https://www.diva-portal.org/smash/get/diva2:1897521/FULLTEXT02.pdf
- https://github.com/admiralakber/physio-rom
- https://www.mdpi.com/1424-8220/25/3/667
- https://www.mdpi.com/1424-8220/23/14/6445