From RGB-D to RGB-Only: Reliability and Clinical Relevance of Markerless Skeletal Tracking for Postural Assessment in Parkinson's Disease.

Although the MP_3D_2 maintained a moderately significant correlation for shoulders and hips (ρ ≥ 0.49), the substantial absolute bias in mean angles (Figures in S1 of Supplementary Materials ) indicates that this model applies a systematic offset across the cohort, rather than failing to capture inter-subject variability, a phenomenon previously noted in healthy subjects [ 62 ].

also in 2,174 other papers