Refined shoulder kinematics via markerless bony landmark detection and acromial 3D shape using an RGB-D camera during hand-cycling
Amedeo Ceglia, Lucas Mulhaupt, Florent Moissenet, Mickël Begon, Lama Séoud
This study develops a markerless method using a depth-sensing camera to accurately estimate upper-limb kinematics during hand-cycling, achieving 5 mm landmark detection errors and 6.3 degree joint angle errors.
bioRxiv · Dec 9, 2025 · PreprintBioengineering