Brief

Researchers have developed a novel pipeline capable of predicting in vivo joint contact forces from monocular video without invasive measurements or subject-specific models. This system utilizes parametric body meshes and a transformer model, incorporating self-supervised video tokens, to accurately estimate forces in hips and knees. The method achieves accuracy comparable to complex musculoskeletal simulations and shows promise for clinical applications like retrospective analysis and at-home rehabilitation tracking. AI