Answers are grounded in this paper and cite their sources.
Physical performance tests such as the 30-s sit-to-stand (30-s STS) test, Timed Up and Go (TUG), standing-balance tasks, the five-times sit-to-stand (5xSTS) test, and gait speed are widely used to quantify functional capacity. This article describes a reproducible validation workflow for a two-dimensional (2D) red-green-blue (RGB) camera-based markerless motion capture (MMC) application. The workflow combines standardized test administration, simultaneous RGB and depth-camera acquisition, independent clinician scoring, an operational tracking-quality rule, independent criterion annotation, test-retest assessment, second-clinician rescoring, and robustness testing under altered acquisition conditions. In a verified pilot implementation involving 15 healthy adults and seven physical-performance tests, 203 of 210 initial planned standard-condition acquisitions before technical replacement were analyzable (96.7%). Application-to-criterion intraclass correlation coefficients (ICCs) ranged from 0.781 to 0.972 across tests; the lower value for side-by-side standing occurred in the setting of restricted between-participant range and ceiling behavior. Stress-condition analyzability was 43/54 (79.6%). These pilot results characterize feasibility, agreement, repeatability, and failure modes; they are not interpreted using a retrospectively imposed binary validation threshold.
No takes yet. Share an insight, caveat, or question.
Yue et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: