Objective: This study aimed to evaluate the agreement and intra-tool reliability between Kinovea (a manual video analysis tool) and a Python-based markerless pose estimation system powered by MediaPipe for deep squat posture biomechanical analysis.Methods: Thirteen asymptomatic participants (8 males, 5 females; 24.23.1 years; BMI30) performed three squat repetitions.Movements were captured using two synchronized HD cameras from frontal and right lateral views.Key indices-peak hip/knee flexion, SHK/HKA angles, and proportional symmetry indices-were analyzed.Statistics included paired t-tests for bias, Cohen's d for effect size, and ICC(3,1) for reliability.Results: No significant differences were found between tools for proportional indices or SHK angles (p0.05).Although HKA2 variance approached significance (p0.051), its absolute mean difference (-2.13) was clinically negligible.Both tools demonstrated excellent relative reliability across all variables (all ICCs0.89).Proportional indices showed the highest consistency (ICC0.956), followed by SHK (ICCs 0.924) and HKA angles (ICCs 0.895).Conclusions: Under controlled 2D conditions, both the MediaPipe-based markerless system and Kinovea demonstrated high agreement and excellent reliability for deep squat analysis.This cross-validation identifies proportional symmetry indices (ICC 0.956) as a robust shared metric.Practically, Kinovea suits small-scale, detail-oriented research requiring expert adjudication.Conversely, the automated markerless system is ideal for large-scale, high-throughput applications like sports screening.Future studies should validate these findings in clinical populations and under 3D conditions to broaden generalizability.
Qiu et al. (2026) studied this question.