Abstract

Postural sway assessment provides key insight into the neural and biomechanical control of balance. Although center of pressure (CoP) trajectories obtained from force plates remain the gold standard, their reliance on specialized infrastructure restricts their broader application. Consequently, kinematics of the head and trunk have emerged as practical surrogates. However, a systematic comparison of their relative validity remains lacking. To address this gap, we first evaluated head and trunk motions recorded via optical motion capture against CoP trajectories during quiet standing under six optic-flow perturbations. Time- and frequency-domain analyses revealed stronger segment-CoP coupling for the trunk than for the head, particularly along the anterior–posterior axis. Feature-level concordance further showed that trunk-derived sway metrics more closely approximated CoP-derived measures. Building on these findings, we next evaluated a minimal, markerless method using a single overhead CCTV camera and a Lucas–Kanade optical flow algorithm. The resulting anterior-posterior trajectories closely matched those derived from motion capture, though medio-lateral estimates were less stable. Collectively, these findings establish the trunk as a biomechanically robust surrogate for CoP and demonstrate the feasibility of low-cost, infrastructure-based video tracking for postural assessment. This approach holds particular promise for scalable use in clinical, telehealth, and community-based settings..