Real time estimation of vertical jump height with a markerless motion capture smartphone app: A proof-of-concept case study

Abstract

The aim of the present proof-of-concept case study was to explore the potential of a novel technology using artificial intelligence techniques to measure countermovement jump height (CMJ-h) in real time. Four hundred jumps were recorded from a single male participant over a period of 24 consecutive weeks, while CMJ-h was simultaneously registered with a force plate and a newly developed version of the My Jump Lab iOS app that used computer vision to measure CMJ-h in real time with the iPhone camera. A very high correlation (r = 0.971, 95% CI = 0.963–0.975) and large agreement (ICC = 0.969, 95% CI = 0.963–0.975) were observed between measurements. Statistically significant, large differences were observed between instruments (mean absolute difference = 0.06 ± 0.01 m, d = 4.4, p < 0.001). However, when using the regression equation between devices to correct the app’s raw data (R² = 0.94, y = 1.0004x – 0.0641), non-significant, trivial differences were observed (mean absolute difference = 0.01 ± 0.008 m, d = 0.1, p = 1.000). Collectively, the findings of this study highlight the potential of this novel artificial intelligence app for the measurement of CMJ-h in real time. However, considering the nature of this investigation, more research is needed to confirm the results observed in a wider population.