Advancing Field-Based Vertical Jump Analysis: Markerless Pose Estimation vs. Force Plates

The countermovement vertical jump (CMJ) is widely used in sports science and rehabilitation to assess lower body power. In controlled laboratory environments, a complex analysis of CMJ performance is usually carried out using motion capture or force plate systems, providing detailed insights into athlete's movement mechanics. While these systems are highly accurate, they are often costly or limited to laboratory settings, making them impractical for widespread or field use. This study aimed to evaluate the accuracy of MMPose, a markerless 2D pose estimation framework, for CMJ analysis by comparing it with force plates. Twelve healthy participants performed five CMJs, with each jump trial simultaneously recorded using force plates and a smartphone camera. Vertical velocity profiles and key temporal variables, including jump phase durations, maximum jump height, vertical velocity, and take-off velocity, were analyzed and compared between the two systems. The statistical methods included a Bland-Altman analysis, correlation coefficients (r), and effect sizes, with consistency and systematic differences assessed using intraclass correlation coefficients (ICC) and paired samples t-tests. The results showed strong agreement (r = 0.992) between the markerless system and force plates, validating MMPose for CMJ analysis. The temporal variables also demonstrated high reliability (ICC > 0.9), with minimal systematic differences and negligible effect sizes for most variables. These findings suggest that the MMPose-based markerless system is a cost-effective and practical alternative for analyzing CMJ performance, particularly in field settings where force plates may be less accessible. This system holds potential for broader applications in sports performance and rehabilitation, enabling more scalable, data-driven movement assessments.