What question did this study set out to answer?

This study evaluates the agreement and reliability between Kinovea and MediaPipe in analyzing deep squat posture.

April 1, 2026Open Access

A Direct Comparison Between Kinovea and a MediaPipe-Based Markerless Pose Estimation System for Deep Squat Posture Analysis : A Reliability and Agreement Study

Puntos clave

This study evaluates the agreement and reliability between Kinovea and MediaPipe in analyzing deep squat posture.
Thirteen asymptomatic participants performed three squat repetitions.
Movements were captured using synchronized HD cameras from frontal and right lateral views.
Key indices like hip/knee flexion and symmetry indices were analyzed using statistical tests.
No significant differences were found in the proportional indices or SHK angles between tools.
HKA variability approached significance, but the mean difference was clinically negligible.
Both tools demonstrated excellent reliability with ICCs over 0.89, with proportional indices achieving the highest ICC of 0.956.

Resumen

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.

Me gusta

Guardar

Ver artículo completo